High Transformer Secondary Neutral Current With No Load

[jjustice]

I will briefly outline the situation regarding my question...
-3 Phase 3 wire supply, ~13400V
-1500KVA transformer, brand new, 13800/4160Y/2400, set on 13460V tap
-energized through a recloser (setup to operate as a switch on a three minute timer)... thus all three phases closed at the same time
-overhead line comes down and goes through a very short underground run up through the pad to the primary. All cables are in individual runs.
-NO cables connected on the secondary... purely open connection.
-There is a secondary ground strap running from X0 to ground.

When the transformer was closed in via the recloser, the controller panel read 140A, which seems like a reasonable amount of inrush for a transformer rated at 64A (full load) on the primary. One minute into closing the readout says 160A. Two minutes into closing the readout says 180A. This levels out and stays at 180A until the three minute timer opens the circuit. This current reading has been verified by a separate clamp on meter as well. This is most definately NOT a typical inrush curve. It appears to be loaded somehow... I am assuming some sort of fault. The last critical piece of information is that the secondary ground strap was glowing cherry red through all this. The core is a 5 limb wound core design. This unit has been pulled from location and is currently in our shop. It has passed the insulation resistance test, turn ratio test, insulation power factor test, winding resistance test, energize test, impedance test and induced potential tests with flying colors. There doesn't appear to be anything wrong with the transformer based on the standard battery of tests that we do here. Everyone is stumped. I was hoping one of you guys could shed some light on the potential cause of such a tremendous neutral to ground voltage that it would overheat the ground strap in that fashion. Having a 5-leg core and a delta primary, it would seem that it is redundantly protected from excessive magnetizing imbalances. Let me know what you guys think! Thanks!

 

[davidbeach]

You've got two grounds, one being the ground strap and the other being something/somewhere else. It would be really helpful if there were a relay event report that showed actual current and voltage waveforms.

 

[davidbeach]

Oh, and why a five leg core if it has a delta primary?

 

[jjustice]

Im not saying there isnt another ground. Ive considered the possibility myself. The problem is none of the testing showed that ground. The site has been rigorously tested... the primary cables have been energized sticking up out of the pad with nothing connected to them and no current flowed... ruling out the primary cabling. The insulation resistance tests done on the transformer showed between 20 and 40 gigaohms on the three tests (LV-G,HV-G,HV-LV).

As for the core selection... a 5-leg 4-loop iron is the easiest way that I can think of to build a wound core. They could have done a three leg evans core, but I have never seen one of those coming from this particular manufacturer. This particular core is done in two halves though. Each ground strap is brought out from the inside layer and insulated, then wrapped around the bottom of the core and set in the frame. essentially grounding the inside of the core. The frame is insulated from the outside layer by both the core insulation and press-board type paper in the frame. Ive also considered the possibility of the core shorting out along the cross section (nullifying that magnetic circuit) causing this problem in the field and then not causing it here. It seems that if it could vibrate enough to go through the ground strap insulation it could also move away from it in transportation as well. I suspect this would cause an imbalance and perhaps result in a high neutral to ground voltage as well and the induced current to flow from the neutral through the ground strap. This seems like a possibility, but not a likelihood.

 

[ijl]

Each ground strap is brought out from the inside layer and insulated, then wrapped around the bottom of the core and set in the frame. essentially grounding the inside of the core.

This does not mean a closed loop around the core?

 

[jjustice]

No, it just means that the inside sheet of each half is tied together so that they are on the same potential. When I say it is 8 piece, I mean its a dual 4-loop design. Picture a core cut down the middle (width) to make laminations of half the original width. The goal of this is increased stacking ease with large cores. When stacked, there is insulation between the two halfs. When grounded on the same layer, they should perform like identical halves. The other pieces are connected with a strip between the outside layers, as is typical of wound core construction. Thus, all outside sheets are at the same potential, as are all inside sheets... and each half performs identically, thus together perform as one.

 

[prc]

Please try low voltage excitation test. ie apply single phase 230 V and measure excitation current in each phase. Do the test from HV side ( line to line) and on LV side (line to X0). Repeat test with X0 earthed and floating. You may be able to find out the fault point.
When was the induced test etc was done? Are you a manufacturer or repairer ?

 

[jjustice]

prc- I am actually in the process of doing this now! I figured this would be the best way to find out which phase was contributing to the fault. I am glad you suggested the same line of testing. I will post results tomorrow. The company I work for is a vendor for the manufacturer. We do sales, service and repair. We do warranty work for the manufacturer on a regular basis, so we have been commissioned to find/fix the problem on this unit.

 

[waross]

At what voltage did you megger the 2400 Volt windings? I have seen a cable pass a 250 Volt megger test in the meg-ohm range but fail with 120 Volts AC. With the megger on the 500 Volt scale, the cable read in the meg-ohms when the megger was cranked slowly but as the speed was increased the needle would abruptly jump from over 1 meg-ohm to zero on the meg-ohm scale.
You may have a similar condition here, given the build-up of current.
Is there anyone else out there who remembers when we had to crank our meggers to start them? Grin

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[rhatcher]

If you picture the circuit through which we are thinking the current is flowing, it leads from the external XO terminal to the ground strap. From there is must be going through the earth ground, to the ground for the frame of the transformer, and then from the frame back to the neutral point (XO) through the internal neutral grounding jumpers.

I am thinking that there is fault in the transformer or that there is a serious difference in potential between the grounding points for the external XO ground strap and the transformer frame. In this case, a transfomer fault actually seems more likely to me. Of course, I could be way wrong...

Have you tried energizing the transformer and measuring the voltage potential between the XO terminal and the transformer frame? Theoretically the internal neutral bonds to the frame should mean that there is no voltage between these points. If you detect a voltage, then there is a fault.

Be certain that if you do this test that you use appropriate measurement equipment. If there is an internal fault, it may be that the voltage difference between these two points exceeds the normal limit for a hand held multimeter.

 

[jjustice]

Results of single phase energizing are as follows:

All phases were energized at full voltage through the primary...

H3-H2@14490V...X0-G=259V
H2-H1@14490V...X0-G=248V
H1-H3@14490V...X0-G=139V

I am going to unground the tank from the supply and try this test again. I believe there is a ground reference in our supply that might be affecting these numbers, but I have not had time to investigate that yet.

 

[Kiribanda]

jjustice,
For me it looks like there is a serious unbalalnce in the low side.How do you have a potential difference between X0 (star point) and the remote ground (G) when the transformer primary is fed by a balanced 3-phase voltage?The moment you connect the X0 (star point) to ground there is definitely a current flow.That is what you have seen by getting the X0 to ground jumper cherry red.In my opinion there is an internal fault in one of the wdgs which should be fully opened and investigate.

 

[jjustice]

Well I thought about this while I was at lunch today. I must comment on our test setup, as I believe the numbers obtained in that test are more of a result of where the neutral was to begin with. At 480V and no load, my test bank shows 265, 198 and 300 volts phase to neutral. It is an open delta setup fed by a rheostat. I believe this is the cause of the strange neutral voltages, in combination with the 14400 delta/480 delta step up bank used to energize the primary. The neutral has to move around to satisfy the magnetizing characteristics for all the transformers in the circuit, so the line currents can balance. Thus, I believe the results of this test at this point are inconclusive. I am going to high-pot the unit now and look down inside as the test is in process to see if I can see anything. I doubt anything will turn up. I am really starting to suspect something else in the system...

 

[jjustice]

The unit passed high pot just fine... primary winding at 34KV and 125mA, secondary winding at 15KV and 75mA.

Let me also answer some old questions that I missed.
-The megger was set on 5KV... electric megger, not hand crank.
-The high voltage induced and applied tests were done in our shop after it was pulled from the pad and traveled 45 miles to our shop.
-Insulation resistance, winding resistance and ratio tests were done in shop and on site, after the problem, before it was pulled.

 

[waross]

If you have an insulation breakdown at a particular voltage you probably won't find it with a 5 kV megger on 34 kV and 15 kV windings.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[jjustice]

The windings are 13.8KV +/- 2.5% and 2.4KV... B.I.L. for the windings are 95 and 60, respectively. 34KV/15KV are the IEEE/ANSI standards for high-pot testing at those levels.

 

[DTR2011]

Have you checked for core grounds?

 

[rhatcher]

jjustice, energize all three phases of the transformer primary and measure the secondary XO-ground (XO-transformer frame) voltage. I guarantee that you will see a voltage that, based on your description of an internal X0-ground bond, should not be there. As to the cause of it, it would have to be an internal fault in the transformer.

 

[Miguel Villarroel]

After all that you've protrayed, one can figure that the problem is not the transformer, the installation shall be causing the trouble

If there was no electricity there would be no internet. Good point, don´t you?

 

[jjustice]

smallgreek... that was my first suspicion... fortunately it did not pan out.

mav... after my final run of tests on friday i was convinced of this also...

So... I solved the problem. We went on site today with a bank of transformers, ready to hook them up and see if the circuit mimicked the performance of the padmount. However, we never got that far. Upon exiting the truck I noticed that the 14400/24940Y-240/480 three phase bank (which we sold them as well) installed on the adjacent pole was hooked up wye. If it werent for myriad mis-information supplied to us, I could have sorted this out on the phone in the first few minutes. According to the man I am apprenticing, this happens all the time. Perhaps I will ask for a picture next time instead of a verbal connection verification. I appreciate all the help an input from all the engineers here at eng-tips.com! Thanks everyone!